Zoom looks to launch next era of custom enterprise AI

Zoom Communications has formed a partnership with Nvidia to deliver “a faster, more secure and customisable” enterprise artificial intelligence (AI) experience that is more private, powerful and personalised to drive enhanced productivity.

The deal will see the AI-first work platform provider’s federated architecture expand to include Nvidia Nemotron open technologies to support AI Companion 3.0 across industries such as finance, healthcare and government.

The framework is expanding to “a next-generation” hybrid language model approach – an AI architecture that can intelligently route queries between Zoom’s proprietary small language models (SLMs), optimised for low latency and quality for a particular skill or tasks, and a fine-tuned large language model (LLM) for complex reasoning. It will also look to accelerate enterprise productivity and collaboration experiences with optimised cost efficiency, quality and latency.

The hybrid machine learning model predicts and generates human language, AI architecture and intelligently links Zoom’s SLMs when responding to queries, optimised privacy and stronger complex reasoning in LLMs. 

The federated architecture is designed to select the best AI model for each task with Nvidia’s advanced AI infrastructure, as well as allow enterprises and government organisations to gain both open and closed model access with stronger efficiency.

Using Nvidia Nemotron allows Zoom’s AI Companion to work across Microsoft 365, Microsoft Teams, Google Workspace, Slack, Salesforce and ServiceNow. Using Nvidia Nemotron reasoning models to enhance decision-making and manage multistep tasks will empower the AI Companion module to reason more deeply, automate workflows and deliver intelligent assistance throughout the enterprise, said Zoom.

Through this approach, Zoom said that AI Companion could use Nvidia’s advanced AI infrastructure, services and software, alongside other models, delivering enhanced capabilities while optimising cost for customers. This includes Zoom’s new 49-billion-parameter LLM, based on Nvidia Nemotron and developed with Nvidia NeMo tools which were “built to achieve optimal balance of speed, cost and accuracy”. Having both open and closed models is seen as delivering greater cost efficiency, faster AI workflows, deeper reasoning capabilities and more efficient collaboration within AI Companion.

Looking further forward, Zoom regards its new initiative as being able to power the next evolution of Zoom’s federated AI architecture, which dynamically incorporates optimal models, including the Llama Nemotron Super-based reasoning model to deliver a balance of accuracy, performance and cost.

“We’ve increased our speed and enhanced lower-cost model decision making using Nvidia GPUs and AI software stack, helping to optimise AI Companion’s core capabilities and enable faster go‑to‑market timelines,” said Zoom chief technology officer X.D. Huangt.

“With the help of Nvidia Nemotron open technologies, we’re accelerating the development of our enterprise retrieval-augmented generation [RAG] capabilities, allowing AI Companion to work seamlessly with Microsoft 365, Microsoft Teams, Google Workspace, Slack, Salesforce and ServiceNow. This partnership allows us to deliver powerful, security-focused and scalable AI experiences to our customers at rapid speed.”

Kari Briski, vice-president of GenAI software at Nvidia, added: “The integration of Nvidia Nemotron into Zoom AI Companion allows enterprises like Nvidia to experience a private, powerful and personalised work environment for enhanced productivity. Our collaboration with Zoom to bring smart and efficient AI reasoning to customers is just beginning.”